A few years ago the debate over New York City’s then new indoor smoking ban had much of the city talking and thinking about when, where and how people smoke. Around that time Professor Nachum Sicherman, attending dance performances, began to notice a lot of dancers outside smoking after the performances. “The contradiction of seeing a person smoke who you would assume puts a high premium on staying healthy puzzled me,” he says.
As a labor economist, Sicherman wondered if smoking could serve as a proxy for time preference, the degree to which a person is oriented to the present or the future. Dancers have relatively short careers, with little prospect of future income, and this suggested to Sicherman that they were perhaps more present-oriented than future-oriented.
Economists are naturally interested in measuring differences in time preference because many economic decisions involve tradeoffs between present and future benefits. “Labor economists,” Sicherman says, “always assume that one of the factors affecting wage changes over time is how much individuals invest in their human capital, in their own job training and education, and that differences in investments are at least partially due to differences in time preference.” The more future-oriented people are, the more likely they are to spend time and money on developing their careers.
But since time preference is inherently unobservable, there is little empirical evidence demonstrating the role that time preference plays in decision making. How do you measure something you can’t see?
“Time preference is not a trait you can observe in a direct way, so we hypothesized that an indirect way to observe it is to assume that, on average, people who smoke place less value on the future than people who do not smoke,” Sicherman explains.
Many studies have determined that smokers earn less than nonsmokers, but Sicherman was interested in how both groups’ wages changed over time. Working with Lalith Munasinghe of Barnard College, Sicherman examined data from the National Longitudinal Surveys of Youth (NLSY), which gathered annually over a 15-year period information about the health, income, employment and education of thousands of respondents.
The researchers hypothesized that if smokers were indeed more present-oriented than future-oriented, they would, on average, see their wages rise more slowly than those of nonsmokers. (The team also conducted a survey at Barnard that showed that among students of all majors dancers were by far the most likely to be smokers.)
Munasinghe and Sicherman focused on the respondents’ first decade in the workforce because it’s during this time that most wage growth related to individual investments in human capital takes place (most people’s wages grow fast early in their careers and then slow over time). The researchers’ primary interest was tracking wages after respondents started their first jobs, and they found a clear correlation between wage growth and smoking. Munasinghe and Sicherman found a 4.7 percent gap (after controlling for a range of family and individual characteristics) between smokers’ and nonsmokers’ first wages (what people were paid at their first full-time job).
Though this result was consistent with many previous findings that smokers earn less than nonsmokers, earlier studies didn’t take wage dynamics into account. Because Munasinghe and Sicherman did consider wage dynamics, their results were more definitive: over the first decade of employment, the difference in wages increased dramatically — nonsmokers’ wages grew to be anywhere from 15 to 40 percent higher than those of smokers. The far-reaching NLSY data allowed the researchers to eliminate variables other than smoking — including sex, age, race, health, schooling, cognitive ability, religion and neighborhood income — that might correlate with the wage changes they observed.
The research has implications for behavioral economics. “Our findings highlight the importance of time preference in individual decision making about the labor market: smokers, presumably because they are more present-oriented, are more likely to self-select into jobs that have lower wage growth and invest less in their own human capital,” Sicherman says. “Social scientists should consider factors that play a role in time preference, and policymakers may in turn want to consider the social costs and benefits influencing time preference in individuals.”
Nachum Sicherman is professor of finance and economics at Columbia Business School. This paper won the Eckstein Prize for best article published in the Eastern Economic Journal for 2005–06.
Read the Research
"Why Do Dancers Smoke? Smoking, Time Preference, and Wage Dynamics"